Crystals
Low-Temperature Photoluminescence Spectroscopy of Solvent-Free PCBM SingleCrystals Giulia Tregnago1†, Michael Wykes2†,Giuseppe M. Paternò1, David Beljonne3 and Franco Cacialli1 †
These authors contributed equally to this work.
1
Department of Physics and Astronomy and London Centre for Nanotechnology, University
College London, Gower Street, London, WC1E 6BT, United Kingdom. 2
Madrid Institute for Advanced Studies, IMDEA Nanoscience, C/Faraday 9, Campus
Cantoblanco, 28049 Madrid, Spain 3
Laboratory for Chemistry of Novel Materials, Université de Mons, 20, Place du Parc, 7000
Mons, Belgium Supporting information
Methods PCBM (Phenyl-C61-butyric acid methyl ester, Sigma Aldrich purity >99%) was dissolved into chlorobenzene at a concentration of 20 mg∙mL-1. The solution was kept under stirring overnight at room temperature. Solvent-free PCBM single-crystals (see Fig. 1b) were prepared as previously reported1: a droplet of PCBM solution was deposited over a fused
silica substrate and the solvent evaporated in air at room temperature. To remove any residual solvent molecules, the sample was dried under vacuum (~10-2 mbar) overnight. We also prepared an amorphous film (~100 nm) spin-coating the PCBM solution in chlorobenzene onto a fused silica substrate at 800 rpm. Furthermore we prepared a polycrystalline film (~800 nm) drop-casting a droplet (100 μL) of PCBM solution over a fused silica substrate and let the solvent to evaporate in air at a constant temperature of 303 K. The powder XRD data are reported in Fig. S1. X-ray diffraction measurements were performed with a Rigaku SmartLab diffractometer (Rigaku, Tokyo, Japan), by using a Kα wavelength emitted by a Cu anode (0.154 nm). The measurements were carried out using a grazing incidence configuration (incidence angle = 0.5°) to reduce the background scattering from the substrate. Variable temperature fluorescence was recorded using a Renishaw Raman System 1000 equipped with a He-Ne laser light source (λex = 632.8 nm) and a CCD detector. To prevent the sample degradation we used low laser power at the crystal surface (~0.1 mW). The Raman spectrometer was configured with a Leica optical microscope equipped with a Nikon SLWD M Plan 40x/0.40 objective. The spot size at the sample surface was ~1-2 µm. Fluorescence spectra in the 4-300 K range were taken placing the sample in a Helium-cooled cryostat (MicrostatHe2, Oxford Instrument). Raman spectra were collected with a Renishaw InVia Raman microscope illuminating the samples with an argon ion laser (488 nm and 514 nm) and a 785 nm laser through a Leica N Plan 50x/0.75 objective. The laser power at the sample surface is 0.1 mW.
Temperature-dependent photoluminescence of the single-crystals
Figure S1 Normalized photoluminescence as a function of temperature (in the range 4-300 K) for a PCBM single-crystal. The spectra are red-shifted compared to those in Fig.1 (see main text).
X-ray diffraction
Figure S2 X-ray diffraction patterns for the polycrystalline (black) and amorphous (red) films
Temperature-dependent photoluminescence of amorphous and polycrystalline films
Figure S3 Fluorescence spectra as a function of temperature for the amorphous (A) and polycrystalline (B) films
Vibronic Coupling Simulations
Unfortunately, the modelling approach applied by Sassara to C602, namely calculating the vibronic envelopes of each pure diabatic excited state (T1g, T2g, Gg) cannot be readily applied to PCBM. In PCBM, the lower symmetry imposes a mixing of states, preventing study of the underlying higher-symmetry diabatic states. The (multiply degenerate) T1g, T2g, and Gg states of C60 are replaced with 10 quasi-degenerate states of mixed character in PCBM, too close in energy to allow for reliable prediction of state ordering and mixing by quantum chemical (QC) methods that can be applied to systems of such size. To confirm this, we simulated the spectrum of C60 including FC and HT effects and in the absence of symmetry constraints (thus allowing for mixing of the diabatic states). Our QC and FCHT calculations were performed using the Gaussian 09 package,3 with the B3LYP functional and 6-31g* basis set. Upon geometry optimization of the lowest excited state, we obtained a lower-symmetry
relaxed geometry and a simulated FC-HT spectrum which bore little resemblance to the experimental and previously simulated spectra of C60 (Fig. S4). This confirms that correctly describing the state mixing in C60 (and hence related fullerenes) is indeed too great a challenge for the TD-DFT-based method used and that this approach should not be adopted for reliable simulation of PCBM PL spectra.
Figure S4 Comparison of simulated vibronic envelopes of C60 PL based on i) Weighted sum of the vibronic envelopes of the pure low-lying diabatic T1g, T2g, Gg excited states developed by Sassara et al. (red curve) and ii) FC-HT vibronic envelope calculated for a single mixedcharacter, low-symmetry lowest excited state obtained by TD-DFT geometry optimization of the lowest excited state in the absence of symmetry constraints (black curve).
Figure S5 Comparison of simulated FC-HT convoluted vibronic envelopes of C60 and PCBM (solid lines) and their underlying individual transitions (stick spectra). Both simulations were based upon a single mixed-character, low-symmetry lowest excited state obtained by TDDFT geometry optimization of the lowest excited state in the absence of symmetry constraints. Though the spectra bear little resemblance to observed spectra, the similarity between the modelled C60 and PCBM spectra support our assumption that we can treat the chemical differences between C60 and PCBM as a small perturbation, and hence adapt Sassara’s model of C60 PL to PCBM.
Figure S6 Comparison of PCBM single-crystal (sample A) 4K PL (red curve) with the simulated spectra of C60 in Ne (black curve) and Ar (blue curve) solid matrices at 4K. The adiabatic origins E00 of both C60 spectra are shown with dashed lines. The C60 simulated spectra are reconstructions of those developed by Sassara et al., which reproduced the equivalent experimental spectra very satisfactorily.
Figure S7. First step in adapting Sassara’s model to PCBM single-crystals. The electronic origin of the C60 modelled spectra in Ne and Ar have been moved to 14200 cm-1 to coincide with the first peak in PBCM PL.
Figure S8. Second step in adapting Sassara’s model to PCBM single-crystals. The weights of the T1g, T2g and Gg states were adjusted by hand to try to best-reproduce relative peak intensities. The band around 13500 cm-1 suggests significant Gg character. T1g, T2g and Gg weights in the summed spectra shown in red are 40%, 20 and 30%, respectively.
Figure S9. Adapting Sassara’s model to include 00 intensity. The 00 line is given an intensity of 200, approximately 2.5 times more intense than the most intense HT false origin (4 t1u vibration coupled to the T1g state with intensity 78.5).2 The 00 of each state has the same FC and JT envelope as the HT false origins of that state.
Figure S10. Tuning the between T1g state and 1 hg vibration to account for the suggestion by Sassara et al that the calculated value of the coupling was likely underestimated due a limitation in the quantum chemical method used. The Huang-Rhys factor was increased from 0.06 to 1.5, improving the relative intensities of the first and second peaks in the spectrum, but falling short of making the second peak more intense than the first.
Figure S11. Comparison of the Gg and Hg vibronic envelopes showing that the Hg state features a peak at 13770 cm-1 (marked with arrows) that is not present in the spectra of any of the other states.
Bibliography 1. Paternó, G.; Warren, A. J.; Spencer, J.; Evans, G.; Sakai, V. G.; Blumberger, J.; Cacialli, F., Micro-focused X-ray diffraction characterization of high-quality 6,6 -phenyl-C61-butyric acid methyl ester single crystals without solvent impurities. J. Mater. Chem. C, 2013; 1, 5619-5623. 2. Sassara, A.; Zerza, G.; Chergui, M.; Negri, F.; Orlandi, G., The visible emission and absorption spectrum of C-60. J. Chem. Phys. 1997, 107, 8731-8741. 3. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09 Revision D.01, Gaussian Inc. Wallingford CT, 2009.
These authors contributed equally to this work.
1
Department of Physics and Astronomy and London Centre for Nanotechnology, University
College London, Gower Street, London, WC1E 6BT, United Kingdom. 2
Madrid Institute for Advanced Studies, IMDEA Nanoscience, C/Faraday 9, Campus
Cantoblanco, 28049 Madrid, Spain 3
Laboratory for Chemistry of Novel Materials, Université de Mons, 20, Place du Parc, 7000
Mons, Belgium Supporting information
Methods PCBM (Phenyl-C61-butyric acid methyl ester, Sigma Aldrich purity >99%) was dissolved into chlorobenzene at a concentration of 20 mg∙mL-1. The solution was kept under stirring overnight at room temperature. Solvent-free PCBM single-crystals (see Fig. 1b) were prepared as previously reported1: a droplet of PCBM solution was deposited over a fused
silica substrate and the solvent evaporated in air at room temperature. To remove any residual solvent molecules, the sample was dried under vacuum (~10-2 mbar) overnight. We also prepared an amorphous film (~100 nm) spin-coating the PCBM solution in chlorobenzene onto a fused silica substrate at 800 rpm. Furthermore we prepared a polycrystalline film (~800 nm) drop-casting a droplet (100 μL) of PCBM solution over a fused silica substrate and let the solvent to evaporate in air at a constant temperature of 303 K. The powder XRD data are reported in Fig. S1. X-ray diffraction measurements were performed with a Rigaku SmartLab diffractometer (Rigaku, Tokyo, Japan), by using a Kα wavelength emitted by a Cu anode (0.154 nm). The measurements were carried out using a grazing incidence configuration (incidence angle = 0.5°) to reduce the background scattering from the substrate. Variable temperature fluorescence was recorded using a Renishaw Raman System 1000 equipped with a He-Ne laser light source (λex = 632.8 nm) and a CCD detector. To prevent the sample degradation we used low laser power at the crystal surface (~0.1 mW). The Raman spectrometer was configured with a Leica optical microscope equipped with a Nikon SLWD M Plan 40x/0.40 objective. The spot size at the sample surface was ~1-2 µm. Fluorescence spectra in the 4-300 K range were taken placing the sample in a Helium-cooled cryostat (MicrostatHe2, Oxford Instrument). Raman spectra were collected with a Renishaw InVia Raman microscope illuminating the samples with an argon ion laser (488 nm and 514 nm) and a 785 nm laser through a Leica N Plan 50x/0.75 objective. The laser power at the sample surface is 0.1 mW.
Temperature-dependent photoluminescence of the single-crystals
Figure S1 Normalized photoluminescence as a function of temperature (in the range 4-300 K) for a PCBM single-crystal. The spectra are red-shifted compared to those in Fig.1 (see main text).
X-ray diffraction
Figure S2 X-ray diffraction patterns for the polycrystalline (black) and amorphous (red) films
Temperature-dependent photoluminescence of amorphous and polycrystalline films
Figure S3 Fluorescence spectra as a function of temperature for the amorphous (A) and polycrystalline (B) films
Vibronic Coupling Simulations
Unfortunately, the modelling approach applied by Sassara to C602, namely calculating the vibronic envelopes of each pure diabatic excited state (T1g, T2g, Gg) cannot be readily applied to PCBM. In PCBM, the lower symmetry imposes a mixing of states, preventing study of the underlying higher-symmetry diabatic states. The (multiply degenerate) T1g, T2g, and Gg states of C60 are replaced with 10 quasi-degenerate states of mixed character in PCBM, too close in energy to allow for reliable prediction of state ordering and mixing by quantum chemical (QC) methods that can be applied to systems of such size. To confirm this, we simulated the spectrum of C60 including FC and HT effects and in the absence of symmetry constraints (thus allowing for mixing of the diabatic states). Our QC and FCHT calculations were performed using the Gaussian 09 package,3 with the B3LYP functional and 6-31g* basis set. Upon geometry optimization of the lowest excited state, we obtained a lower-symmetry
relaxed geometry and a simulated FC-HT spectrum which bore little resemblance to the experimental and previously simulated spectra of C60 (Fig. S4). This confirms that correctly describing the state mixing in C60 (and hence related fullerenes) is indeed too great a challenge for the TD-DFT-based method used and that this approach should not be adopted for reliable simulation of PCBM PL spectra.
Figure S4 Comparison of simulated vibronic envelopes of C60 PL based on i) Weighted sum of the vibronic envelopes of the pure low-lying diabatic T1g, T2g, Gg excited states developed by Sassara et al. (red curve) and ii) FC-HT vibronic envelope calculated for a single mixedcharacter, low-symmetry lowest excited state obtained by TD-DFT geometry optimization of the lowest excited state in the absence of symmetry constraints (black curve).
Figure S5 Comparison of simulated FC-HT convoluted vibronic envelopes of C60 and PCBM (solid lines) and their underlying individual transitions (stick spectra). Both simulations were based upon a single mixed-character, low-symmetry lowest excited state obtained by TDDFT geometry optimization of the lowest excited state in the absence of symmetry constraints. Though the spectra bear little resemblance to observed spectra, the similarity between the modelled C60 and PCBM spectra support our assumption that we can treat the chemical differences between C60 and PCBM as a small perturbation, and hence adapt Sassara’s model of C60 PL to PCBM.
Figure S6 Comparison of PCBM single-crystal (sample A) 4K PL (red curve) with the simulated spectra of C60 in Ne (black curve) and Ar (blue curve) solid matrices at 4K. The adiabatic origins E00 of both C60 spectra are shown with dashed lines. The C60 simulated spectra are reconstructions of those developed by Sassara et al., which reproduced the equivalent experimental spectra very satisfactorily.
Figure S7. First step in adapting Sassara’s model to PCBM single-crystals. The electronic origin of the C60 modelled spectra in Ne and Ar have been moved to 14200 cm-1 to coincide with the first peak in PBCM PL.
Figure S8. Second step in adapting Sassara’s model to PCBM single-crystals. The weights of the T1g, T2g and Gg states were adjusted by hand to try to best-reproduce relative peak intensities. The band around 13500 cm-1 suggests significant Gg character. T1g, T2g and Gg weights in the summed spectra shown in red are 40%, 20 and 30%, respectively.
Figure S9. Adapting Sassara’s model to include 00 intensity. The 00 line is given an intensity of 200, approximately 2.5 times more intense than the most intense HT false origin (4 t1u vibration coupled to the T1g state with intensity 78.5).2 The 00 of each state has the same FC and JT envelope as the HT false origins of that state.
Figure S10. Tuning the between T1g state and 1 hg vibration to account for the suggestion by Sassara et al that the calculated value of the coupling was likely underestimated due a limitation in the quantum chemical method used. The Huang-Rhys factor was increased from 0.06 to 1.5, improving the relative intensities of the first and second peaks in the spectrum, but falling short of making the second peak more intense than the first.
Figure S11. Comparison of the Gg and Hg vibronic envelopes showing that the Hg state features a peak at 13770 cm-1 (marked with arrows) that is not present in the spectra of any of the other states.
Bibliography 1. Paternó, G.; Warren, A. J.; Spencer, J.; Evans, G.; Sakai, V. G.; Blumberger, J.; Cacialli, F., Micro-focused X-ray diffraction characterization of high-quality 6,6 -phenyl-C61-butyric acid methyl ester single crystals without solvent impurities. J. Mater. Chem. C, 2013; 1, 5619-5623. 2. Sassara, A.; Zerza, G.; Chergui, M.; Negri, F.; Orlandi, G., The visible emission and absorption spectrum of C-60. J. Chem. Phys. 1997, 107, 8731-8741. 3. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09 Revision D.01, Gaussian Inc. Wallingford CT, 2009.
